Research On Public Opinion Of Automobile Enterprises Based On Text Analysis

With the continuous improvement of people’s living standards,automobile has become a popular means of travel.The competition among auto companies is becoming increasingly fierce,and the public opinion of auto companies is an important part of the competition.The fermentation of public opinion events also reflects the importance of network public opinion analysis.With the help of text sentiment classification,enterprises can understand the trend of network public opinion and the center of public opinion,so as to correct the negative impact of public opinion and maintain the brand image of enterprises.This paper will take the vehicle enterprise network public opinion data as the research object,and the public opinion emotion classification and public opinion topic extraction are studied.In the sentiment classification analysis of public opinion,the collected data is firstly cleaned.Then,according to the text data,the sentiment classification model of public opinion is established from three perspectives.The first is to build SVM,which is the basic model of this paper.The second is to build the deep learning model,including CNN,LSTM.The third is to improve the algorithm according to the advantages and disadvantages of the deep learning model,including Bi LSTM,CNN-LSTM and LSTM-CNN.Finally,the performance of the models is evaluated by the accuracy and macro index.The experimental results show that the public opinion emotion classification model based on CNN-LSTM is the best.In the analysis of topic extraction of public opinion,LDA topic model is established to extract topics from the data,and the K of topic number is determined by the perplexity of the model.Through topic extraction of positive public opinion and negative public opinion,it is found that the public’s concerns mainly focus on automobile quality,automobile cost performance,automobile production and marketing.According to the results of the analysis,we can put forward guiding suggestions for enterprises to improve the quality of goods and improve the corporate image.